Font Size: a A A

Research On Key Technologies Of Multi-source Guiding Information Fusion

Posted on:2017-05-12Degree:DoctorType:Dissertation
Country:ChinaCandidate:X HuangFull Text:PDF
GTID:1108330482991287Subject:Mechanical and electrical engineering
Abstract/Summary:PDF Full Text Request
With the development of microelectronics technology, computer science and sensor technology, multi-source information fusion has become a new area of research. And multi-source information fusion is widely applied in the military and civilian field. On the basic of the characteristic of multi-source information fusion, some of the key technologies of multi-source guiding information fusion in the framework of its model are analyzed in this thesis, particularly the target recognition fusion and the threat assessment technology. Some of algorithms for target recognition and threat assessment are proposed for an electro-optical guiding and tracking system. The main contributions of this dissertation are as follows:1. The basic theory of information fusion and its model are introduced. And the research and development status of target recognition and threat assessment are summarized. The advantages and disadvantages of several existing algorithms for target recognition and threat assessment are analyzed. The main technical means of this dissertation are identified.2. In view of the characteristics of interval-valued intuitionistic fuzzy sets and grey relational analysis, a new method for target recognition in the electro-optical guiding and tracking system based on interval-valued intuitionistic fuzzy grey correlation is proposed. This method obtains characteristic parameters of unknown targets according to the multi-source guiding information. And interval intuitionistic fuzzy numbers are used to denote the membership and non-membership between each characteristic parameter and each target category, in order to form a target recognition matrix. After extracting the positive and negative ideal recognition strategies, the grey correlation theory is used to analyze the recognition matrix to construct the grey correlation order of each target category. The experimental results show that the recognition error of the target recognition algorithm based on interval valued intuitionistic fuzzy grey correlation is less than that of the single method. And the issue of combination explosion can be avoided.3. In view of the drawbacks of traditional methods in comparing real-valued intuitionistic fuzzy values or interval-valued intuitionistic fuzzy values, a new method named synthesis function method is proposed, which is applied in the analysis of target recognition matrics. In this method, the score function and the accuracy function are primarily taken into account to construct the synthesis function with a variable weight coefficient, which depends on the risk sensitivity of a decision maker. The synthesis function method can effectively improve the defects caused by the maximization of net profit in the traditional method, and provide an estimate for the distance between the two fuzzy values.4. In view of some characteristics of particle swarm optimization algorithm(PSO) and BP neural network, the PSO algorithm is modified. And then the target threat assessment algorithm based on BP neural network optimized by modified PSO algorithm(MPSO-BP) is proposed to apply in the electro-optical guiding and tracking system. This method uses the PSO algorithm to optimize the initial weights and thresholds of BP neural network. And during the process of PSO optimization, in order to avoid particle population fast convergence effection, mutation operator and optimization for several parameters are introduced in PSO. The experimental results show that the prediction error of MPSO-BP network is less than that of BP network and PSO-BP network, and it can solve the problem of small sample training and avoid the local extremum of neural network.5. Accoring to some characteristics of glowworm swarm optimization(GSO), krill herd algorithm(KH) and support vector machine(SVM), the models for target threat assessment based on SVM optimized by GSO algorithm(MGSO-SVM) and SVM optimized by KH algorithm(KH-SVM) are established respectively. And the algorithm based on each model is proposed. The two optimized SVM networks are applied in the electro-optical guiding and tracking system to estimate the target threat degree better. The experimental results show that the prediction error of GSO-SVM or KH-SVM is obviously less than KH-SVM.
Keywords/Search Tags:multi-source guiding information, electro-optical guiding and tracking system, information fusion, target recognition, threat assessment, interval-valued intuitionistic fuzzy, grey correlation, optimization algorithm
PDF Full Text Request
Related items